4

Click here to load reader

Chlorine monofluoride. Reactions with sulfur oxides

Embed Size (px)

Citation preview

Page 1: Chlorine monofluoride. Reactions with sulfur oxides

Vod. 9, No. 2, February 1970 CHLORINE MONOFLUORIDE 3 11

the smaller deviation from ideality. A smooth curve fits the data for the ternary mixtures except for mixtures 1 and 5, Table I. The calculated heats of solution for these solvents appear to be rather low. It should be noted that if the indicated error in solubility a t the high temperature is in one direction and the error a t the low end of the range has the opposite sign, a con- siderable change in the apparent heat of solution would result.

The values shown in Figure 7 for the heat of solution of CeF3 in the two LiF-BeF2 mixtures listed in Table I are higher than those calculated for LiF-BeF2 solvents with higher BeF2 concentrations5 and are much greater than more recent values'O for the heats of solution of NdF3 and SmF3 in LiF-BeF2 (66-34 mol %), 11.5 and 10.5 kcal/mol, respectively. Consequently, the shape of the curve in the low-ThF4 end of Figure 7 is not firmly established.

(10) F. A. Doss, F. F. Blankenship, and J. H. ShatTer, Report ORNL-4119, Oak Ridge National Laboratory, Oak Ridge, Tenn., July 1967, p 144.

0 0.1 0 2 0 3 0 4 0 5 0 6 07 0 8 0 9 10 p -ThFq/ lThFq+BeFpi

Figure 7.-Correlation of apparent heat of solution with solvent composition.

Acknowledgment.-The authors thank J. H. Shaffer and coworkers for supplying purified solvent composi- tions and members of the ORNL Analytical Chemistry Division under the supervision of W. R. Laing and L. J. Brady for performing wet chemical analyses.

CONTRIBUTION FROM ROCKETDYNE, A DIVISION OF KORTH AMERICAN 91304 ROCKWELL CORPORATION, CANOGA PARK, CALIFORNIA

Chlorine Monofluoride. Reactions with Sulfur Oxides

BY CARL J. SCHACK AND RICHARD D. WILSON

Received July 2 , 1969

The reactions of C1F with SOF,, SOz, and SO3 are reported. tively, in excellent yields. interaction of ClF with HSOlF and HS03Cl is also reported.

These reactions produce SOF4, ClS02F, and C1OSOzF, respec- New characterization data for ClOSOzF are presented and some of its reactions are described. The

Introduction Chlorine monofluoride has been shown to react in a

variety of ways. For example, it can function as either a Lewis acid' forming the C1F2- anion or a Lewis base2 forming the ClpF+ cation. In addition, i t can act as a chlorinating a fluorinating agent,b or a chloro- fluorinating material. The latter form of reaction can be either catalyzed6*' or uncatalyzed.8 Our interest in the chemistry of ClF prompted an examination of the course of its reaction with a series of sulfur-oxygen- containing compounds.

Experimental Section Apparatus.-Reactions were conducted in stainless steel

cylinders and the products were separated and purified using a stainless steel-Teflon FEP (polyfluoroethylenepropylene) vac- uum line. Pressures were measured by means of a Heise Bour- don-tube type pressure gauge. Infrared spectra were taken on

(1) K. 0. Christe and J. P. Guertin, Inovg . Chem., 4, 1785 (1965). (2) K. 0. Christeand W. Sawodny, ibid. , 8, 212 (1969). (3) C. J. Schack, ib id . , 6, 1938 (1967). (4) D. Pilipovich and C. J. Schack, ibid., 7, 386 (1968). (5) K. Dehnicke, Angev . Chem., 76, 385 (1964). ( 6 ) C. J. Schack and W. Maya, J . A m . Chem. Soc., 91, 2902 (1969). (7) C. J. Schack, R. D. Wilson, J. S. Muirhead, and S. N. Cohz, ibid., 91,

(8) W. K. R. Musgrave, Adwnn. Fluorine Chenz., 1, 18 (1960). 2907 (1969).

Perkin-Elmer Infracords 137 and 337 using 5-cm path length stainless steel or Kel-F cells fitted with AgCl windows. The 19F nmr spectra were obtained using a Varian Associates high-resolu- tion nmr spectrometer operating a t 56.4 Mc. Samples were sealed in Kel-F tubes and CFCla was employed as the external standard. Mass spectral data were obtained with a CEC 21-103C mass spectrometer with the inlet system operating a t room tem- perature and employing an ionization voltage of 70 V. A Hanovia 100-W utility lamp (Catalog No. 30620) was used for the photolysis.

Materials.-Chlorine monofluoride was prepared by heating an equimolar mixture of Clz and ClFI to 150' for several hours in a stainless steel cylinder. Thionyl fluoride was made from thionyl chloride and NaF in acetonitrile.e Sulfur dioxide (Matheson Co.) and sulfur trioxide (Baker and Adamson) were commercial materials and were purified by removal of impurities not con- densable at -196". Fluorosulfonic (Baker and Adamson) and chlorosulfonic (Eastman) acids were used as received. Cesium fluoride and potassium fluoride were fused and then powdered in a drybox prior to use. In all cases the chlorine monofluoride was purified immediately before use by vacuum fractionation through a U trap cooled to - 142'.

Reaction of C1F with Thionyl Fluoride.-Thionyl fluoride (102 cm3, 4.55 mmol) and C1F (211 cm3, 9.42 mmol) were suc- cessively condensed at - 196' into a prepassivated 30-1111 cylin- der. The cylinder was then allowed to warm from -78" to

Soc., 84, 4275 (1962). (e) F. A. Fawcett, C. W. Tullock, and D. D. Coffmann, J . A m . Chem.

Page 2: Chlorine monofluoride. Reactions with sulfur oxides

312 CARL J. SCIIACK AND RICIIARD r). WIT,SC)N Inorganic; Chemistry

TABLE I MASS SPECTRUM OF CISOzF

Re1 Re1 Iona intens m / e Ion intens

120 37C1SOzF + 15.54 64 SOz+ 8 .38

118 ClSOzF+ 41.21 51 SF+ 2.43 101 37c1s02+ 3.04 48 SO+ 11.89 99 C1SOz' 8.11 37 37C1+ 2.43 85 %OzF+ 5.40 35 C1+ 7.19 83 SOzF' 100.00 32 S+ 14.59 69 3 4 s 0 ~ + 2.70 16 O + 3 .78

67 SOF+ 37.16

C134SOaF +

37C1S +

CIS +

a Sulfur-32 and chlorine-35 unless otherwise indicated.

ambient temperature over several hours and left a t that tem- perature overnight. Subsequently the reaction products were separated by fractional condensation in U traps cooled a t - 126, -142 and -196". The warmest trap contained Clz and a trace of SO2FC1 while the coldest trap contained unreacted CIF and minor amounts of SOFz, SOR, and SOzF2. The intermediate temperature fraction was found to be SOFI contaminated with a considerable amount of Cln. The latter was removed by reac- tion with mercury at ambient temperature and in this way pure SOF4 (90.8 cm3, 4.05 mmol) was obtained in 90% yields based on the amount of SOF2 used. The purity and identity of the products were based on vapor pressure and infrared and mass spectral measurements.lof" Reactions using equimolar quanti- ties of SOFz and CIF were also conducted at ambient tempera- ture and a t -78". These experiments also resulted in the formation of SOP4 in quite high yield based on one SOFa being formed from one SOF:! and two ClF's.

Reaction of C1F with Sulfur Dioxide.-Sulfur dioxide (127 cm3, 5.67 mmol) and CIF (162 cm3, 7.23 mmol) were succes- sively condensed into a 30-ml prepassivated cylinder cooled to -196'. The cylinder was then warmed from -78' to ambient temperature over several hours. Products of the reaction were separated by fractional condensation a t -78, -112, -142, and -196'. hTo unreacted SO2 was found and only small amounts of Clz and SOzF2 were noted along with unreacted CIF. Sulfuryl chloride fluoride (125.3 cm3, 5.59 mmol) was isolated in the trap cooled a t -112" in 997, yield based on the amount of SOz used. The product was identified by its infrared12 and mass spectra (Table I).

Reaction of ClF with Sulfur Trioxide.-In a typical reaction, SO3 (191 cm3, 8.53 mmol) and excess CIF were coadensed into a prepassivated 30-ml cylinder a t - 196". The cylinder was slowly warmed to room temperature (>2-hr warm-up) and fractiona- tion was initiated through traps cooled to -78, -95, -142, and -196". The -142 and -196" fractions consisted of unre- acted CIF with small amounts of Clz and S02Fz. The other frac- tions contained the pale yellow product ClOSOzF in 90+70 yield based on the amount of so3 used. Continued pumping on the -78' fraction a t that temperature resulted in the distillation of nearly all material from that fractionation. Measurements indi- cated the material to be pure ClOSOzF: vapor density: obsd, 135.8 g/mol; theoret, 134.5 g/mol; IgF nmr: lit.,13 singlet a t -33.9 ppm; obsd, singlet a t -33.3 ppm with no other detect- able resonances. In addition, the mass spectrum of the product agreed with that reported13 for ClOSOZF. The following tem- perature-vapor pressure data were observed ("C, mm): -31.0,

302. The vapor pressure-temperature relation is described by the equation log P,, = 8.0951 - (1650.6/T°K). The normal

19; -23.8,30; -10.0,66; 0.0, 113; 5.9, 152; 12.7,209; 20.8,

(10) F. B. Dudley, G. H. Cady, and D. F. Eggers, Jr.. J. Am. Chem.

(11) P. L. Goggin, H. L. Roberts, and L. A. Woodward, Tuans. Favaday

(12) T. Birchall and R. J. Gillespie, Specluochim. Acta, 22, 681 (lQG,?). (13) W. P. Gilbreath and G. H. Cady, Inoug. Chem., 2, 49G (1963).

Soc., 78, 1553 (1956).

SOL., 57, 1877 (1961).

boiling point calculated from the equation is 43.4', with a heat of vaporization of 7.55 kcal/mol and a Trouton constant of 23.8.

The infrared spectrum of ClOSOzF has not been reported pre- sumably because of the rapid attack of the compound on cell window materials. For example with AgCl an obscuring back- group can form almost a t once. This background has an inira- red spectrum which is nearly identical with that reported1* for AgSOBF. However, CIOSOzF that had been fractionated as above resulted in only slow background formation. Further- more samples that had been distilled a t -95' resulted in almost no background buildup. This indicates that this reaction is catalyzed by an impurity. The infrared spectrum of ClOSOzF is shown in Figure 1 and the principal bands are a t 1480 (s), 1248 (s), 855 (s), 832 (sh), 703 (m), 571 (m), and 529 (w) cm-'.

:IAVELINGlh i1:iICRO'lSl

Figure 1.-Infrared spectrum of CISOBF at 80 and 8 mni

Reaction of ClOSOzF with CsF.--A 10-ml prepassivated cylin- der was loaded with CsF (0.58 g, 3.82 mmol) and evacuated and then ClOSOzF (68.4 cm3, 3.05 mmol) was added at -196'. The cylinder was allowed to warm to approximately 0' for 30 min before i t was recooled t o -142' and the material volatile a t that temperature was pumped off and trapped a t -196". The volatiles were found by infrared examination t o be pure C1F (16.8 cm3). After rewarming to room temperature for 10 hr an additional 24.4 cm3 of CIF was removed from the cylinder after recooling to -142'. h subsequent 3-day period a t room tem- perature gave an additional 7.2 cm3 of CIF and the unreacted ClOSOzF (20.2 cm8, 0.90 mmol) was then recovered. The total amount of CIF formed (48.4 cm3, 2.16 mmol) represented a 71% yield based on the amount of ClOSOpF used. The white solid left in the cylinder was weighed (0.75 g) and the weight gain was found to agree with that expected for a mixture of 2.16 mmol of CsSOaF and 1.66 mmol of CsF (0.75 g). The presence of CsS03F in the solid was confirmed by comparing its infrared spectrum to that reportedI4 for CsSOgF. Potassium fluoride also reacted with ClOSOzF to generate ClF.

Photolysis of ClOSOzF.-A 100-nil Pyrex bulb equipped with a Fisher-Porter Teflon needle valve was loaded with ClOSOzF (94 cma, 4.2 mmol). While maintaining the bulb a t approxi- mately -78", the contents were irradiated for several hours with uv light. The products of the photolysis were separated by fractional condensation in traps cooled to -78 and -196'. The latter contained practically pure Clz and the former had only Sz06Fz (43 cm3, 1.9 mmol) as determined by its infrared spectrum and vapor pressure.15

Reaction of C1F with HS03F and HSOsCl.-Quantitative data on the reaction of fluorosulfonic acid and C1F were not obtained. However, it was shown that the reaction was not particularly fast since an equimolar mixture of the two materials still had 157, unreacted C1F after 2 hr a t room temperature. In addi- tion, appreciable amounts of Ch, FC102, and Clz were formed along with the expected HF. The formation of ClOSOzF by the reaction was indirectly determined by treating the fraction trapped a t -78' with KF. This treatment resulted in a 30% yield of ClF based on the original I-ISOgF used. An equimolar quantity of HS03C1 and C1F (6.30 mmol each) reacted a t room

(14) D. W. A. Sharp, J . Chem. Soc., 3761 (1967). (15) I?. B. Dudley and G. H. Cady, J . Ant . Chem. Soc., 79, 813 (1957)

Page 3: Chlorine monofluoride. Reactions with sulfur oxides

Vol. 9, No. 2, February 1970 CHLORINE MONOFLUORIDE 313

temperature to form a nearly quantitative yield of Cl2 (139 cm3, 6.20 mmol). The relatively nonvolatile liquid product left was identified as HSOaF with some unreacted HSOlCl by record- ing the infrared spectrum as a film between AgCl plates and comparing it to the literature spectra.16

Results and Discussion The reaction of thionyl fluoride and C1F

SOFz + 2C1F + SOFa + Clz

occurred readily to give thionyl tetrafluoride in high yield and provides a useful alternate synthesis of that compound when FZ is not available. Variation of the reactant ratios and the reaction temperatures (- 78 to 25") did not alter the course of the reaction. The observed, facile oxidative fluorination of thionyl fluoride using C1F is somewhat surprising when com- pared to the more stringent conditions described1°*17 where elemental fluorine is the fluorinating agent. It is possible that an initial chlorofluorination occurs to give a product such as SOF3Cl which then is rapidly fluo- rinated to SOF4. However, efforts to verify the forma- tion of such an intermediate by using lower tempera- tures and limited amounts of ClF were not successful. Additional fluorination or chlorofluorination of SOFk was not expected or encountered since it does not occur except in the presence of certain alkali metal fluorides.?

Sulfur dioxide and chlorine monofluoride react rapidly and completely to give a nearly quantitative yield of sulfuryl chloride fluoride. Additional ClF did not react with ClSOzF under the present conditions. The reaction

SO2 + CIF ---t ClSOzF

represents a more convenient synthesis of this material since previous preparations, using SbF3, l8 NH4F,lg or K F in CH3CNZ0 to fluorinate SOZClz, give lower yields and are more tedious. The exclusive chlorofluorination observed with SO2 is markedly different from the ex- clusive fluorination noted with SOFz and ClF. No obvious explanation for this difference in reactivity is apparent.

Sulfur trioxide and ClF reacted readily to form chlorine fluorosulfate in high yield

SOs + CIF ----f ClOSOzF

Only one other method of preparing ClOSOzF has been reported13 and that involved the reaction of Clz and peroxydisulfuryl difluoride. The present reaction offers an attractive alternate since the use of Sz06Fz may be hazardous if the potentially explosive impurity, fluorine fluorosulfate, 21 is not completely removed.

While the ClF reactions with SOFz or SO2 involved only attack on the sulfur central atom, in the case of SO3 addition across one of the S=O double bonds occurred. This is the first example of the uncatalyzed addition of C1F to an S=O group. A previously reported addi-

(16) R. Savoie and P. A. Giguere, Can. J. Chem., 42, 277 (1964). (17) W. Kwasnik in "Handbook of Preparative Inorganic Chemistry,"

Vol. I, G. Brauer, Ed, Academic Press Inc., New York, N. Y., 1963, p 171. (le) H. S. Booth and C. V. Hermann, J. A m . Chsm. Soc., 68, 63 (1936). (19) M. M. Woyski, ibid. , 72,919 (1950). (20) J. K. Ruff, Inoug. Chem., 4, 567 (1965). (21) G. H. Cady, Chem. Eng. News, 44 (E), 40 (1966).

tion' to an S=O group (ie., the conversion of SOFI to SF60C1) required a catalyst. However, both additions produce compounds containing the ClOSF grouping as would be expected from a consideration of the polarity of the starting materials. Here too no additional fluorination occurred with excess ClF, a result which appears to be typical of high-valent chloroxy-substi- tuted species such as CF30C16 and SFSOC~.'

Measurements of several properties of CIOSOzF (vapor density, 19F nmr, and mass spectra) agreed well with the values originally reported.13 However, all the samples we prepared differed with regard to vapor pressure, our measured values being somewhat higher a t all temperatures. This discrepancy may be due to the fact that the original work was carried out in glass with which ClOSOzF reacts slowly whereas the present work was conducted in a stainless steel-Teflon F E P system with which ClOSOzF did not show any sign of reaction after proper passivation.

The infrared spectrum of ClOSOzF which has not been reported previously is shown in Figure 1. The absorptiohs found were quite similar to those of other covalent f l~o rosu l f a t e s '~~~~ and showed the characteris- tic absorptions for SO2 (1480, 1248, and 571 cm-l) and SF (855 cm-l) groups. The band observed a t 703 cm-l occurs ia the range6.' characteristic for a C1-0 stretching vibration. A detailed analysis of the vibra- tional spectrum of ClOSOzF is in progress.z3

Chlorine fluorosulfate reacted with K F and CsF as shown by

This reaction represents the first example of the genera- tion of a chlorine fluoride from a covalent noninter- halogen compound. The liberation of only C1F indicates a preferred attack of the fluoride ion on the positively polarized chlorine atom resulting in the rupture of the C1-0 bond. Comparably substituted compounds such as CF30Cl and SF60C1 do not react with alkali metal fluorides under similar condition^.^^^ Thus, it appears that the stability of the anion formed on breaking the C1-0 bond (SOSF- vs. CF30- or SFSO-) is important in determining whether these compounds are readily susceptible to fluoride ion attack.24 The related reaction of fluorine fluorosulfateZ6 follows a different stoichiometry and has been interpreted as proceeding through a primary attack on the sulfur. This leads to the generation of SOzFz by cleavage of the S-0 bond.

Photolysis of ClOSOzF a t low temperature was found to give peroxydisulfuryl difluoride readily

ClOSOzF + M F + MSOaF + ClF (M = K, CS)

hu 2ClOSOzF --f Cla + FSOzOOSOzF

This reaction is the reverse of the original synthesis13 of

(22) F. B. Dudley, G. H. Cady, and D. F. Eggers, Jr., J. Am. Chem. SOC.,

(23) K. 0. Christe, C. J. Schack, E. C. Curtis, and W. Sawodny, tu be

(24) Chlorine nitrste, which might be expected to undergo this reaction to unpublished results this laboratory and

78, 290 (1956).

submitted for publication.

give the nitrate ion, does not do su: ref 3.

(25) J. K. Ruff and M. Lustig, IvLoug. Chem., S, 1422 (1964).

Page 4: Chlorine monofluoride. Reactions with sulfur oxides

314 G. E. RYSCIIK&WITSCII AND J. W. M’IGGINS Inorgunic Chemistry

C10S02F. Cleavage of the C1-0 bond by ultraviolet radiation has been generally observed6!’ to produce the respective peroxide in good yield.

The reactions of HN03 or HNFz with ClF have been demonstrated3t4 to be fast and efficient methods of forming the respective chlorine-substituted compounds CINOI and C1NF2. However, fluorosulfonic acid and ClF reacted more slowly and with appreciable formation of the by-products Cl,, FC102, and (2102. These compounds are indicative of a ClF-H20 reactionjZ6 the latter being an impurity in the commercial HSO3F used. Nevertheless, C10S02F was obtained in a t least

(2G) R. Bougon, M. Carles, and J. Aubeit, Coinpl R e n d , C266, 178 (1967).

-___

30% yield. Chlorosulfonic acid and an equivalent amount of C1F reacted as shown by

HSOaCl + ClF ----t HS03F + Clz

rather than giving the unknown compound, chlorine chlorosulfate. Thus, only fluorination occurred through halogen exchange and no evidence for chlorina- tion of the OH group was obtained.

Acknowledgment.-The authors gratefully acknowl- edge support for this work by the Office of Kava1 Re- search, Power Branch. We are also indebted to Drs. D. Pilipovich and K. 0. Christe for helpful discussion

CONTRIBUTIOX FROM IHE DEPARTMENT OF CHEMIS~KY, UXIVERSITY OF FLORIDA, GAI~ESVILLE, FLORIDA 32601

Hydrogen Elimination in Dimethylamine-Borane

BY G. E. RYSCHKEWITSCH AND J. W. WIGGINS

Receioed August 18, 1969

The elimination of hydrogen from dimethylamine-borane \vas studied a t 100’ by observing the hydrogen isotope distribu- tion in the products resulting from reactants with different hydrogen isotopes on nitrogen and/or boron. The results clearly eliminate the possibility of a unimolecular elimination and favor a bimolecular reaction. Different deuterium isotope effects on the elimination rates were deduced, depending on whether the nitrogen atom or the boron atom was substituted (kN-H/kN-D = 3.5; kB-H/kB-D = 1.2). Various possible isotopic exchange reactions which could have complicated the study were investigated.

Introduction Dimethylamine-borane and other borane adducts of

primary and secondary amines have been reported to lose hydrogen a t elevated temperature to yield amino- boranes, borazine, and redistribution products or substituted boranes.

The first step in these reactions has been demon- strated to follow the stoichiometry expressed by

RzSHBHa + R2NBH2 + Hz (1)

It would be of interest to determine whether this reaction proceeds via a unimolecular elimination, as is the case in a number of organic systems, or whether elimination occurs between more than one molecule. The normal experimental approach, where the rate law would be established from concentration measurements, seemed inappropriate since the amine-boranes are not very volatile nor are they, as mixtures of the products, readily susceptible to quantitative analysis. Moreover, the aminoborane products may dimerize, trimerize, or redistribute to give bis(aminoboranes) and amino- diboranes. Therefore, the problem was approached by labeling boron and/or nitrogen bonds with deuterium and following the reaction by isotopic analysis of the noncondensable products.

(1) E. Wieberg, A. Bolz, and P. Buchheit, Z. Anoig. C h e m . , 266, 286 (1948).

(2) H. I. Schlesinger, D. M. Ritter, a n d A. B. Burg, J . A m . CIaeiia. Soc., GO, 1296 (1938).

(3) A. B. Burg aud C. L. Kandolph, Jr . , ihzd., 73, 958 (1851).

Discussion It was proposed to answer two basic questions in this

study. (1) Is the hydrogen elimination reaction in dimethylamine-borane unimolecular or not? (2) LlThat is the magnitude of the hydrogen isotope effect?

The first question was answered by the first two experiments in Table I, where mixtures of about equal proportions of undeuterated and completely deuterated compounds were heated or where mixtures of (CH3)2- NHBD3 and (CH3)sNDBHy were heated. Intramolec- ular hydrogen elimination should, in the first instance, produce only H2 and D2, whereas in the second instance, only HD should be formed. Incomplete deuteration might in either set introduce a small quantity of the missing isotopic variety of molecular hydrogen. On the other hand, intermolecular hydrogen elimination should produce all three hydrogen species in propor-

TABLE I

DIMETHPLAMINE-BORANES DECOMPOSITION O F r ) E U T E R A T E D

Amt of species, yo --70 products---

Isotopic species mmol reaction Dn HD I3 2

(CHa)2NDBDa 0.82 31.3 11.0 47.7 41.3 (CHa)nNHBHa 0.77 (CHs)2KDBI-Ia 0.50 32.5 12.9 46 .5 4 0 . 0 (CHa)zNHBDa 0 ,48 (CHa)gNUBI)y 1 .05 38 83.1 16 .9 . . . (CHa)pNL)BD3 1.47 3 . 9 60.5 39.5 . . ,